NOTES
3008
tally by the usual procedure. The dielectric constants were €150 200.1, €250 182.4, €350 174.3, and €450 167.1 a t the temperatures indicated and were those reported by Leader and Gormley.' The values of viscosities a t the different temperatures were 7160 = 0.0199, 7250 = 0.0165, 7350 = 0.0142, and 7450 = 0.0123 The values of the limiting transport numbers, t+O, of K+, at various temperatures, obtained from the plot of the Longsworth function against concentration, are also given in Table I under the concentration heading 0.00 M .
Results and Discussion As can be seen from Table I, the variations in the transport number, t+, of K ion in N-methylformamide are very much similar to those in formamide3 and in K-methyla~etamide;~i.e., t+ decreases with increase in concentration and increases with increase in teniperature. The limiting transport number of potassium ion, i.e., t+*(K+)! increases from 0.4945 at 15" to 5290 at 45". This behavior is opposite to that in aqueous solutions in which t+"(K+)decreases slightly with rise in temperature. Recently, Gill9 has reported similar behavior of the cationic transport number of KN03 in liquid ammonia in which t+"(K+)has been found to decrease with rise in temperature in the range -65 to -45", although t + ' ( ~ ~ + )t +, " ( L i + ) , and ~ + " ( N H , + ) increase with rise in temperature as t + " ( ~ + )(KBr) does in K-methylformamide. It may be noted that t + " ( ~ +increases ) beyond 0.5 with rise in temperature. This is a little unusual although such values of the transport number have been reported in the literature.1° Ionic Mobilities The limiting cationic transport numbers at different temperatures can be used to calculate the ionic mobilities from the available electrolytic conductance data6 in this solvent and making use of the usual Kohlrausch law of the independent migration of ions. The ionic conductivities at infinite dilution of some ions thus obtained are given in Table 11. Since no ionic conductance data are available jn the literature, it is not possible to verify the values given in Table 11. However, they seem to be reasonable if they are compared with the corresponding values in formaniide and in K-methylacetamide. Since the tetrahedral structure present in water is missing in these solvents, the structure-breaking effect of the larger ions like K + is also missing in these solvents and SO all the ions behave normally. I n view of the lack of sufficient and appropriate electrolytic, and hence the ionic, conductance data in this solvent, it is not possible The Journal o j Physical Chemistry
Table 11: Mobilities of Some Ions in N-Methylformamide a t Different Temperatures Ionic mobility at-------
Ion
Na + K+ CS
+
(Cz&)aN
+
Picrate-
c1-
BrI-
150
25'
17.59 18.05 19.90 21.52 11.28 16.88 18.46 19.38
21.56 22.13 24.39 26.20 13.08 19.70 21.56 22.76
350
450
...
...
27.62
32.15
. . I
...
... . . I
... ...
... ...
25.38
28.60
...
...
to make a reasonable estimate of the solvation of ions as was done in NMA.4 This aspect will be examined later when the required data are available.
Acknowledgnient. A grant from the Society of Sigma Xi and RESA Research Fund of the United Stat,es of America for the purchase of the solvents (NMA, NAIF, and NJIP) is very much appreciated. ~
~~~
~~~~~
~
~
(7) G . R. Leader and J. F. Gormley, J . Am. Chem. SOC.,73, 5731 (1951). (8) R. Gopal and S.A. Rizvi, J . Indian Chem. SOC.,43, 179 (1966). (9) J. B. Gill, J . Chem. Soc., 5730 (1965). (10) B. E. Conway, "Electrochemical Data," Elsevier Publishing Co., London, 1952, p 166.
Photooxidation of Perfluoroethyl Iodide and Perfluoro-n-propyl Iodide'" by Dana Marsh and Julian Heicklen'b Aerospace Corporation, El Segundo, California (Received February 15, 1966)
In earlier reports, the oxidation of CF? and CFClz3 radicals was examined. In the former case, the only carbon-containing product was CF20, whereas in the latter case only CFClO was found. In the presence of HI, the results were unchanged. In this work, we extend the oxidation studies to C2Fs and n-C3F7 radicals in order to determine the products of reaction and the nature of the intermediates. In both cases, the major product is CF20. (1) (a) This work was supported by the U. S.Air Force under Contract No. AF 04(695)-669. (b) To whom requests for reprints should he sent. (2) J. Heicklen, J . Phys. Chem., 70, 112 (1966). (3) D.G, Marsh and J. Heicklen, ibid., 69, 4410 (1965).
NOTES
3009
I n the CzF51system, it is produced with a quantum yield of about 2.0. CF3CF0 is also produced, but is only l/40 as important. With H I present, the oxidation is drastically modified, and the ROz intermediate must live at least sec.
Experimental Section Perfluoroethyl iodide and perfluoro-n-propyl iodide were obtained from Peninsular ChemResearch, Inc. Gas chromatographic analysis showed no impurities in the CzFJ but one major impurity of about 7% in the n-C3F7I. The CzFJ was used after degassing a t -196", but the C3Fd was purified by trap-to-trap fractionation, and that fraction volatile at - 63.5" but condensable at -126" used. I n this way, the impurity level was reduced but not eliminated. Furthermore, the violet color of the impure material disappeared. Matheson research grade 0 2 and anhydrous H I (degassed at - 196") were used. Infrared analyses were performed in situ in a PerkinElmer Nodel 13 Universal spectrometer. A T-shaped cell was used, having 11.5-cm infrared and 10.7-cm ultraviolet path lengths. The ultraviolet light entered the stem of the T through a quartz window. The top of the T had sodium fluoride windows at each end and was situated in the infrared beam. Radiation was from a Hanovia-type SH, U-shaped mercury lamp and passed through a Corning 0-53 glass to remove wavelengths below 2800 A before entering the cell. The effective radiation was primarily at 3130 A with the line at 3020 A also playing some role. The ultraviolet absorption coefficients for HI, C2F51,and C3F71were determined at 3020 and 3130A on a Cary Model 15 spectrophotometer. For longer wavelengths (Le., 3340 A and above) there was no absorption. Absolute quantum yields were measured by comparison with the CFzO produced from photolysis of CF31-02 mixtures where @(CF20)= 1.0. Results and Discussion The photolysis of C2FJ or C3F71yields no products, even for extended exposures. As in all alkyl iodides, the indicated mechanism is RI
+ hv
+I
4R
2R +Rz R+I+RI 21
+ SI +
I2
However, the Iz is a powerful radical very quickly the reaction is inhibited by R
+ Iz +R I + I
With CzF5I-O2 mixtures, the products found were CFZO, CF&FO, and SiF4. Iodine is also formed and perhaps Fz or FzO, though they would not be detected by infrared analysis. The rate of SiF, formation increases with exposure time, indicating secondary decomposition of some product. However, the CF3CFO and CFzO grow linearly with tjime; their quantum yields are listed in Table I . The CZFJ and Oz Table I: Photooxidation of C2FJ (2' = 25", lo = 35 =k 5 p/min) (CZFJ).
(Od,
mm
mm
Q(CFz0)
Q(CF3CFO)
11 10 31 28 28 100 102 100 315 300 300
10 300 10 33 265 10 127 300 10 315 362
2.1
0.072 0.054 0.040 0.072
2.0 1.4 2.7 1.6 1.7 2.1 1.9
0.035
... 2.2 4.8 Av 2 . 0 f 0 . 3
0.069 0.051 0.043 0.056 0.024 0.069 __ 0.053 rt 0.013
pressures were varied from 10 to 300 mm, but the quantum yields were unaffected. They are @(CF20)= 2.0 and @(CF3CFO) = 0.05. The mechanism that most easily explains the results is CzFj
+
0 2
CzFsOz +CFzO
CzFjOz
(6)
+ CFsO
(7)
+ FO
(8)
CzFj02 + CF3CFO
where k7/ks is about 20. The CFIO formed from (7) may have sufficient energy to decompose immediately
CF30*
CF20
+F
(9)
+ Fz
(10)
or it may form CF20 by 2CF30 +2CFz0
With C3F7I-O2 mixtures, the results are more difficult to interpret because of the presence of the impurity. The products observed were the same as in the CzF51-02 photolysis. Initially, the unpurified C3F71 was irradiated and CFzO was produced with a quantum yield of from 10 to 30. Purikation of the C3F71by fractionation reduced the impurity level considerably, but did not eliminate the impurities. Photolysis of mixtures of 0 2 and the purified C3Fd showed the initial large @(CFZO),but the rate of CF20 producVolume 70,Number 9 September 1966
NOTES
3010
tion rapidly dropped to a smaller constant value as the CFzO approached about 1% of the C3F7I. Presumably, the initial reaction is similar to that for CZFS radicals C3F7
f Oz
C3F70z
(11)
I n some manner, the C3F702radical must yield primarily CFzO. How this occurs is not clear at all. It is of interest to know whether the RO2 radical has a measurable lifetime. To that end, we did a number of experiments with 5 to 40 mm of HI added. Two series of experiments were done, one with 280 mm of CzFBIand 300 mm of 02,and the other with 100 mm of C3F71and 300 mm of 02. Hydrogen iodide had no effect at all on CF3 or CFClz o x i d a t i ~ n . ~However, ,~ with C2Fj and C3F7, very marked changes occurred in the oxidation. The results, the same for both series, showed no trend with H I pressure, thus indicating that the initially formed perfluoroalkyl radical was conipletely scavenged by 0 2 . The results can be summarized as follows. First, a red depo:it was formed. Second, a(CF20) was reduced at least 5- to 20-fold. Third, @(CF&FO) may have increased to 0.13 to 0.30 for CzF6radicals as based on the 9.05-p band of CF3CF0. The analysis is difficult because of the H I absorption in this region.
F Fourth, the characteristic C-C=O band at 5.31 p was the most intense band and was formed at the same rate in all experiments with HI. The CF3CFO can account for only 25 to 50% of the 5.31-12 band in the C3F71series. but perhaps all of it in the CZFSIseries. At least for the C3F71series, another product probably was formed There can be no doubt that H I interferes with the oxidation. Consequently, the ROz sec. radical must have a lifetime in excess of
Acknotdedgment. The authors wish to thank IIrs, Barbara Peer for assistance with the manuscript.
Heats of Transport of the Rare Gases
in a Rubber Membrane‘
by Riirion Y. Bearman and Richard J. Bearman Department of Chemistry, University of Kansas, Lawrence, Kansas (Receized February 21, 1966)
Revised values for the heats of transport, &*, of the rare gases He, Xe, Ar, Kr, and Xe in rubber are presented. I n the ease of helium, it is shown that, within The Journal of Phgsical Chemistry
experimental error, there is no pressure dependence of the heat of transport in the pressure range 25-65 cm. The apparent &* varies slightly with temperature difference across the membrane. Nevertheless, departures from linearity in the local phenomenological equations seem to be negligible even though the temperature gradients range up to 900”/in.
Revised Heats of Transport (Q*) In an earlier article, Bearman2 reported the heats of transport (or “transfer”), &*, of the rare gases in a rubber membrane. His values were obtained from thermoosmosis experiments where, at steady state -(&*/R)(l/Tc - ~ / T H ) (1) Here, H and C refer to the hot and cold side, respectively, T is absolute temperature, p is pressure, and R is the gas constant. His measurements were subject to several errors of unknown magnitude arising chiefly from uncertainties in the pressure and temperature. With the use of strain gauge pressure transducers and several thermocouple probes, we have constructed an improved apparatus, fully described elsewhere,3 , 4 in which the errors have been greatly lowered. I n Table I, we present revised values for the heats of transport together with a comparison with the earlier results. The error estimates take into account the observed irreproducibility and also errors arising from temperature and pressure measurements. The irreproducibility is caused mostly by leakage, degassing, and adsorption in the system3 over the long course of the runs, which last days or sometimes weeks. A pure gum rubber membrane 0.0325 to 0.0350 cm in thickness was used for our measurements reported here. In (PH/Pc)-
=
Pressure and Temperature Difference Dependence of Q* Some additional measurements were made with helium. In Table 11, me show that, within the experimental error, the heat of transport at constant mean temperature and temperature difference is independent of mean gas pressure in the system in the range from 25 to 65 cm. From this, we conclude that the He-He interactions in the membrane play little role ( 1 ) This research was supported in its early stages by a grant from the National Science Foundation. We wish to make acknowledgment to the donors of the Petroleum Research Fund, administered by the American Chemical Society, for partial support of this research in its later phases. (2) R. J. Bearman, J . Phys. Chem., 61, 708 (1957). (3) M. Y. Bearman, Dissertation, University of Kansas, 1965. Available from University Microfilms, Ann Arbor, Mich. (4) M. Y . Bearman and R. J. Bearman, J . A p p l . Polymer Sci., 10, 773 (1966).
